Restorations in the form of dental prosthetics may be necessary to address partially or wholly edentulous conditions. Traditionally, such restorations have been performed by forming a model by providing an impression of the affected areas of a patient's mouth, developing a stone model from the impression, and fabricating a customized prosthetic device on the stone model. The process is cumbersome and requires excessive intrusion into the patient's mouth. However, the stone model provides enough accuracy to produce prosthetics that minimize stress and interference with the edentulous area.
Recently, intra-oral scanning (IOS) has emerged as a preferred dental impression technique for conventional (tooth-borne) and implant dentistry. IOS typically involves using a handheld scanner having optical sensors to capture a three-dimensional dataset of the area of interest. The resulting dataset may be used for constructing a model for preparing patient specific prosthetics. An example of using such datasets to construct a model may be found in U.S. Patent Publication No. 2011-0183289, filed on Dec. 7, 2007, titled “Method For Manufacturing Dental Implant Components.” The IOS process offers a very efficient and cost-effective means by which to acquire and transmit anatomic data for purposes of forming a prosthetic. While the accuracy of IOS has been proven to be sufficient for single tooth restorations and short-span multiple tooth segments, it is often contraindicated for scanning larger edentulous segments such as a full arch area scan or potentially smaller segments which are “highly” edentulous.
There are many potential contributing factors for the difficulty of applying IOS to full arch restorations. For example, small adjacent site-to-site errors, while having minimal impact on single tooth or short-span multiple tooth segments, may accumulate where the resulting error throughout the full arch is unacceptable.
While IOS is robust when scanning well defined landmarks (i.e., teeth vs. tissue), large homogeneous areas needed for full arch restoration are problematic. As an arch is scanned, if there are homogeneous segments, especially large ones, these landmarks are vague and, therefore, cannot be interpreted as accurately. The teeth serve as robust landmarks in a scan of an arch, but soft-tissue surfaces between segments of the arch such as the mouth surfaces and the tongue are homogenous surfaces and are therefore difficult to scan accurately. The connecting area(s) such as the tongue or the roof of the mouth are essentially seen as “oceans” of homogeneous surfaces in the scan dataset in that these homogeneous surfaces are difficult to distinguish from each other because they all appear the same in the scanned dataset.
The geometry acquired for the cross-arch connecting geometry (i.e., the tongue or the palate) covers a relatively large area, but only a small portion of the data within this area is scanned. This may lead to cross-arch error and/or full arch distortion and is often most visible when assessing the posterior segments of the resulting model, as these zones are adjacent to the greatest area of “digital dead space” (or the space not scanned). For example, FIG. 1A shows a full arch area 100 which includes an arch 102 having a number of teeth 104 on two segments 106 and 108. In this example, edentulous areas between the teeth 104 on the arch 102 require the application of a dental restoration process. A cross-arch geometric connecting area 110 separates the two segments 106 and 108 of the arch 102. In order to form proper restorative devices, such as a bridge for the arch 102 the distance between the two segments 106 and 108 must be accurately determined.
Scanning the connecting area 110 has limited effectiveness in determining accurate dimensions because the connecting area 110 does not have any distinct features. The connecting geometry area 110 is relatively non-defined (or vague). While a scan of this area eliminates the digital dead space, the quality of the data does not provide for a sufficiently precise digital acquisition and subsequent reconstruction of the dental anatomy of the arch area 100. Such errors are magnified at the end of the segments 106 and 108 due to the geometry of the segments 106 and 108 in relation to the front of the arch 102. For example, a cumulative error of over 180μ for this posterior cross-arch span connecting area 110 shown in FIG. 1A in a resulting model would be much greater than the tolerance allowed to passively seat a full-arch denture supporting bar framework. While the distortion may be small, the clinical relevance of this error is significant, preventing the proper fabrication of the restorative device.
FIG. 1B shows a full arch area 150 which includes an arch 152 with a full edentulous condition with two segments 156 and 158. In this example, edentulous areas on the arch 152 require the application of a dental restoration process. In this example, a series of implants have already been implanted in the arch 102 in preparation for modeling of the arch area 150. Each of the implants 170 has a gingival healing abutment 170 that extends through the soft tissue. A cross-arch geometric connecting area 160 separates the two segments 156 and 158 of the arch 152. In order to form proper restorative devices, such as a bridge for the arch 152 the distance between and around the two segments 156 and 158 must be accurately determined.
FIG. 2A shows a control model 202 formed by a 3Shape laser scan of a cast of the arch area 150 and implants 170 shown in FIG. 1B. The control model 202 is very accurate since it is prepared by scanning the cast produced from a mold taken from the area of interest. Such a larger scan system is more accurate than the handheld scanners used for the IOS techniques because of the differences in the associated algorithms required for acquiring the data and reconstructing the 3-dimensional datasets. FIG. 2B shows a model 204 that is manufactured using a scan dataset from the arch area 150 taken by known IOS techniques. As explained above, the homogeneity of the connecting area 160 results in dimensional inconsistencies between the model 204 and the actual arch area 150. FIG. 2C shows the scan model 204 in FIG. 2B overlaying the control model 202 in FIG. 2A. As shown in FIG. 2C, shaded areas 210 represent distortions between the actual dimensions of the arch area 150 represented by the control model 202 produced by casting and the model 204 produced by known intra-oral scanning techniques. As shown in FIG. 2C, the distortions 210 occur throughout the entire arch, but are greatest on the ends of the segments of the model 204 because of the inaccuracies in determining the dimensions of the connecting geometry between and around the segments. Such inaccuracies may result in positive stretching where the segments of the model are wider than those of the actual arch area. The positive stretching may be seen by the arrows labeled by “EE” in FIG. 2C. The inaccuracies may also result in a model which suffers from negative stretching where the segments are narrower than the actual arch areas. The resulting models therefore are not useful in the restorative process since the resulting prosthesis devices will not interface correctly with the actual arch area.
One proposed solution has been to spray the connecting geometry area with a coating in order to help establish scannable features within the connecting area. The arch and the connecting area are then scanned and a resultant dataset is produced. However, the spraying technique still results in inaccurate scans because the features of the connecting area such as the tongue, assuming that they contain geometry which is distinguishable enough to provide robust data, may move from the location captured during the scan.
Thus, a need exists to improve the accuracy of known intra-oral scanning to enable reliable full arch scanning There is a need to calibrate an intra-oral scan dataset with known dimensions to improve the accuracy of the scanned dataset. There is a further need to perform real-time error correction on a scan dataset in the process of acquisition of the scanned data points.